Related Experiment Video
Updated: May 8, 2026
Cutoff Value of Phase Angle by Bioelectrical Impedance Analysis at Admission as a Prognostic Factor in Patients with Acute Heart Failure
Published on: June 10, 2025
Predictive models for mortality readmission events and cardiovascular complications in patients with COPD: a
Ilektra M Papazoglou1, Hassan Abbas1, Patrick Murphy2,3
1School of Population Health and Environmental Sciences, King's College London, London, UK.
Insights
Machine learning (ML) models enhance prediction for chronic obstructive pulmonary disease (COPD) readmissions but not mortality. Statistical models perform comparably for COPD mortality, and ML models show generalizability issues due to overfitting.
Area of Science:
- Pulmonary Medicine
- Biostatistics
- Health Informatics
Background:
- Chronic obstructive pulmonary disease (COPD) presents a significant global health challenge, linked to high mortality, hospital readmissions, and cardiovascular disease (CVD) complications.
- Predictive models, encompassing statistical and machine learning (ML) techniques, are crucial for risk stratification and informed clinical decision-making in COPD management.
- This review critically evaluates the performance and generalizability of existing predictive models for COPD-related outcomes.
Purpose of the Study:
- To systematically review and assess the performance and generalizability of statistical and machine learning (ML) predictive models for chronic obstructive pulmonary disease (COPD) outcomes.
- To compare the effectiveness of ML versus statistical models in predicting COPD mortality, readmissions, and cardiovascular disease (CVD) complications.
- To identify limitations and suggest future research directions for improving predictive modeling in COPD care.
Main Methods:
- A systematic literature search was conducted across EMBASE, MEDLINE, and PubMed for studies published since 2015 that evaluated predictive models for COPD-related outcomes.
- Studies were screened based on predefined criteria, and model performance was synthesized using meta-analysis, calculating pooled area under the curve (AUC) values for different model types.
- The Prediction model Risk Of Bias ASsessment Tool (PROBAST) was employed to assess the risk of bias in the included studies.
Main Results:
- Out of 3,488 screened records, 37 studies met the inclusion criteria, focusing on mortality (20), readmissions (14), and CVD complications (6).
- Statistical models achieved a pooled AUC of 0.787. For mortality, statistical models performed comparably or better than ML models (AUC 0.801 vs. 0.760). ML models showed superior performance for readmissions (AUC 0.812 vs. 0.758).
- CVD outcomes had a pooled AUC of 0.810. External validation frequently diminished ML model performance, indicating potential overfitting and generalizability concerns.
Conclusions:
- Machine learning models demonstrate improved prediction for COPD readmissions but do not offer a consistent advantage over statistical models for mortality prediction.
- ML models exhibit generalizability challenges, primarily due to overfitting, which impacts their reliability in real-world clinical settings.
- Future research should prioritize real-world validation and explore hybrid modeling approaches to enhance the interpretability and clinical applicability of predictive tools in COPD management.
Background:
COPD is a major global health burden, associated with high rates of mortality, readmissions and cardiovascular disease (CVD) complications. Predictive models, including statistical and machine learning (ML) approaches, have been developed to support risk stratification and clinical decision-making. This review assesses their performance and generalisability.
Methods:
A systematic search of EMBASE, MEDLINE and PubMed identified studies published since 2015 evaluating predictive models for COPD-related outcomes. Studies were screened using predefined criteria, and model performance was synthesised via meta-analysis. Pooled area under the curve (AUC) values were calculated for each model type. Risk of bias was assessed using the Prediction model Risk Of Bias ASsessment Tool (PROBAST).
Results:
Of 3 488 records screened, 37 studies met inclusion criteria: 20 focused on mortality, 14 on readmissions and six on CVD complications. Statistical models had a pooled AUC of 0.787 (95% CI 0.755-0.816). For mortality, statistical models outperformed or matched ML models (AUC 0.801 versus 0.760; p=0.1195), while ML models outperformed statistical approaches for readmissions (AUC 0.812 versus 0.758; p=0.4423). CVD outcomes showed a pooled AUC of 0.810 (95% CI 0.749-0.859). External validation often reduced ML model performance, raising concerns about overfitting.
Conclusions:
ML models improve readmission prediction but offer no consistent advantage for mortality, where statistical models perform similarly. ML models face generalisability challenges due to overfitting. Future work should emphasise real-world validation and hybrid approaches to enhance interpretability and clinical applicability in COPD care.
Related Concept Videos
01:25Chronic Obstructive Pulmonary Disease-III: Symptoms and Complications.
Symptoms of COPD can be classified as primary or systemic. Primary symptoms relate to reduced airflow, while systemic or extrapulmonary symptoms relate to COPD's broader impact on the body.
Primary Symptoms of COPD:
01:29Chronic Obstructive Pulmonary Disease-V: Management
Smoking Cessation
01:27Chronic Obstructive Pulmonary Disease-IV: Assessement and Diagnostic Studies
Medical History
01:30Chronic Obstructive Pulmonary Disease-V: Nursing Management
Assessment
01:24Chronic Obstructive Pulmonary Disease
Smoking is a primary risk factor for COPD, with over 80% of patients having a history of it. Patients typically experience progressive dyspnea or labored breathing, frequent coughing, and recurrent pulmonary infections. Many eventually succumb to respiratory failure, characterized by...
01:20Chronic Obstructive Pulmonary Disease-I: Introduction